Acronym Index + Engineering Doctrine

Licensing & Disclaimer

This document is licensed for internal strategic planning, educational use, and advisory reference only. Redistribution, resale, derivative publication, or use for training machine learning systems without explicit written permission is prohibited.

This material is provided for informational purposes only and does not constitute legal, medical, or regulatory advice. All clinical decisions remain the sole responsibility of licensed healthcare professionals.

How to Use This Playbook

This operational technical manual is written for AI software engineers building clinical-grade systems, with enough context for healthcare stakeholders to understand what the engineering team is doing and why.

If you are designing architecture or defining system boundaries: start with Engineering Doctrine and Reference Architecture.
If you are preparing a study or deployment and need traceability: use Data Plane, Model Plane, and Release Engineering.
If you are facing questions from compliance, partners, or regulators: use Security/Integrity, Change Control, Monitoring, and Regulatory-Aligned SDLC.

In plain language: This is a practical manual for building clinical AI systems that can be trusted, audited, updated safely, and deployed in real clinical workflows.

Quick Reference

Use this index to orient quickly.

Architecture boundaries are contracts. Clients render; the platform decides what runs and records what happened.
Lineage is mandatory. Inputs are immutable; datasets are registered; manifests define what was used.
Models are releases, not weights. Every behavior change creates a new model_version.
Evidence is reproducible. Every metric ties to eval_suite_version, dataset_hash, and metric_spec_version.